On Mon, May 31, 2010 at 04:11:25PM -0500, Matt Domsch wrote:
Doh. This is a result of using tmpfs for the buildroots. There are
a
few such failures, which I'll rebuild in the next run w/o using tmpfs.
Odd, as using tmpfs should have started overflowing into swap, which
doesn't appear to have happened.
The way the tmpfs plugin for mock on F12 is written and the way it was
used in the rebuild, it will only provide a filesystem with half the RAM
size as available space. The only way to change this on F12 is to
manually patch /usr/lib/python2.6/site-packages/mock/plugins/tmpfs.py
and add "-o", "size=Xg" to mountCmd to provide X gigabytes of space.
Regards
Till